Upload 6 files

Dockerfile

@@ -22,6 +22,8 @@ EXPOSE 7860
 # Set environment variables
 ENV CREWAI_TELEMETRY_ENABLED=false
 ENV CREWAI_STORAGE_DIR=/tmp/crewai
+ENV CREWAI_DB_PATH=/tmp/crewai/crewai.db
+ENV CREWAI_MEMORY_ENABLED=false
 ENV HF_HOME=/tmp/huggingface_cache
 ENV TRANSFORMERS_CACHE=/tmp/huggingface_cache
 ENV HF_HUB_CACHE=/tmp/huggingface_cache
@@ -32,6 +34,7 @@ ENV GRADIO_SERVER_PORT=7860
 RUN mkdir -p /tmp/crewai
 RUN mkdir -p /tmp/huggingface_cache
 RUN chmod 777 /tmp/huggingface_cache
+RUN chmod 777 /tmp/crewai
 
 # Command to run the Gradio app
 CMD ["python", "app.py"]

agents.py

@@ -4,6 +4,8 @@ import os
 # Set CrewAI storage directory to something writable BEFORE any imports
 os.environ["CREWAI_STORAGE_DIR"] = "/tmp/crewai"
 os.environ["CREWAI_TELEMETRY_ENABLED"] = "false"
+os.environ["CREWAI_DB_PATH"] = "/tmp/crewai/crewai.db"
+os.environ["CREWAI_MEMORY_ENABLED"] = "false"
 
 from utils import gemini_llm
 from crewai import Agent

app.py

@@ -4,6 +4,8 @@ import os
 # Set CrewAI storage directory to something writable BEFORE any imports
 os.environ["CREWAI_STORAGE_DIR"] = "/tmp/crewai"
 os.environ["CREWAI_TELEMETRY_ENABLED"] = "false"
+os.environ["CREWAI_DB_PATH"] = "/tmp/crewai/crewai.db"
+os.environ["CREWAI_MEMORY_ENABLED"] = "false"
 
 # Create the directory if it doesn't exist
 os.makedirs("/tmp/crewai", exist_ok=True)

main.py

@@ -4,6 +4,8 @@ import os
 # Set CrewAI storage directory to something writable BEFORE any imports
 os.environ["CREWAI_STORAGE_DIR"] = "/tmp/crewai"
 os.environ["CREWAI_TELEMETRY_ENABLED"] = "false"
+os.environ["CREWAI_DB_PATH"] = "/tmp/crewai/crewai.db"
+os.environ["CREWAI_MEMORY_ENABLED"] = "false"
 
 from crewai import Crew, Process
 from agents import finance_knowledge_agent, market_news_agent, stock_analysis_agent, response_refiner_agent

tasks.py

@@ -4,6 +4,8 @@ import os
 # Set CrewAI storage directory to something writable BEFORE any imports
 os.environ["CREWAI_STORAGE_DIR"] = "/tmp/crewai"
 os.environ["CREWAI_TELEMETRY_ENABLED"] = "false"
+os.environ["CREWAI_DB_PATH"] = "/tmp/crewai/crewai.db"
+os.environ["CREWAI_MEMORY_ENABLED"] = "false"
 
 from utils import search_qdrant, search_news, get_stock_data
 from crewai import Task

utils.py

@@ -5,6 +5,8 @@ import os
 # Set CrewAI storage directory to something writable BEFORE any imports
 os.environ["CREWAI_STORAGE_DIR"] = "/tmp/crewai"
 os.environ["CREWAI_TELEMETRY_ENABLED"] = "false"
+os.environ["CREWAI_DB_PATH"] = "/tmp/crewai/crewai.db"
+os.environ["CREWAI_MEMORY_ENABLED"] = "false"
 
 # Set HuggingFace cache directories
 os.environ["HF_HOME"] = "/tmp/huggingface_cache"
@@ -67,19 +69,13 @@ except Exception as e:
         embeddings = SimpleEmbeddings()
         print("⚠️ Using SimpleEmbeddings as final fallback - RAG functionality will be limited")
 
-# Connect to the existing Qdrant collection with error handling
-try:
-    qdrant = QdrantVectorStore.from_existing_collection(
-        embedding=embeddings,
-        url=QDRANT_URL,
-        api_key=QDRANT_API_KEY,
-        collection_name=COLLECTION_NAME
-    )
-    print("✅ Successfully connected to Qdrant")
-except Exception as e:
-    print(f"❌ Failed to connect to Qdrant: {e}")
-    print("⚠️ RAG functionality will be limited - using web search only")
-    qdrant = None
+# Connect to the existing Qdrant collection
+qdrant = QdrantVectorStore.from_existing_collection(
+    embedding=embeddings,
+    url=QDRANT_URL,
+    api_key=QDRANT_API_KEY,
+    collection_name=COLLECTION_NAME
+)
 
 # Initialize Mistral LLM
 mistral_llm = LLM(model="mistral/mistral-large-latest", api_key=MISTRAL_API_KEY, temperature=0.7)
@@ -91,16 +87,11 @@ gemini_llm = LLM(model="gemini/gemini-2.0-flash", api_key=GEMINI_API_KEY, temper
 @lru_cache(maxsize=100)
 def search_qdrant(query, top_k=3):
     """Search Qdrant for relevant documents."""
-    if qdrant is None:
-        print("⚠️ Qdrant not available, returning empty results")
-        return []
-    
     try:
         retriever = qdrant.as_retriever(search_type="similarity", search_kwargs={"k": top_k})
         results = retriever.invoke(query)
         return [{"text": doc.page_content, "source": doc.metadata.get("source", "Unknown")} for doc in results]
-    except Exception as e:
-        print(f"❌ Error searching Qdrant: {e}")
+    except Exception:
         return []
 
 def search_news(query, max_results=5):